[FFmpeg-devel] Possible leak in RTSP to disk copy writing

Carl Eugen Hoyos cehoyos
Mon Mar 15 13:13:07 CET 2010


Stas Oskin <stas.oskin <at> gmail.com> writes:

> First, did you miss this message?
> > http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/96248/focus=96250
> >
> Well, back then I did a really stupid testing procedure by Ctrl-C FFMPEG

If you don't understand a part of my message, it would be a good idea to either
try to guess what I meant or ask (no, this project is not called FFMPEG).

> before letting it clean-up the memory, which I admitted in follow-up email.

Yes, it seems _I_ missed _your_ message;-)

> I did retested it recently, and found out that while all memory indeed
> released in the end, during the run it just growing.

> > Note that "leak" == valgrind complains (and "valgrind does not complain" ==
> > no leak).
> >
> Well, this probably has several points of view, one of them that if even if
> the program releases all correctly, but still eventually runs out of memory
> while running, it can be considered as a leak.

No.

[...]

> > Are you sure you are not describing the properties of writing the iso
> > container?
> >
> This is the exact command I'm using - just dumping the frames to disk.

To rephrase: Are you sure the iso container allows to "just dump the frames"
(no, I don't know, I am just guessing).

Carl Eugen




More information about the ffmpeg-devel mailing list